Dynamic Planning for Obstacle Avoidance of Crawler Based on Gaussian Model

In order to solve the problems of obstacle avoidance planning and trajectory optimization unmanned platform in hilly orchards, this paper proposes a dynamic algorithm based on Gaussian probability model predict behavior. The technical route method theoretical derivation-algorithm design-simulation analysis-real vehicle testing are both adopted. Firstly, global reference path orchard is optimized quadratic programming. Secondly, intentions surrounding pedestrians other obstacles estimated model. Subsequently, grid map created Frenet coordinate system. Then, paths dynamically planned obtain discrete optimal paths. Finally, quintic polynomial curve utilized connect sampling points generate with continuous curvature. analysis by ROS/Rviz simulation indicates that time decreases 57.97%, number sampled nodes 79.57%, curvature smoother compared RRT algorithm. A field test conducted road. With model, can effectively behavioral state obstacles. autonomously plan continuous, smooth safe driving shorter period time, avoiding reducing impact steering-in-place operating efficiency. Through research, behavior prediction proposed for crawler chassis orchards programming, obtained satisfies kinematic constraints chassis, has good safety, better real-time enforceability. Additionally, comprehensive set software hardware solutions spontaneous static presented.